LOSS OF WATER 43 
should be made perfectly air-tight by means of a piece of 
rubber tubing, about i inch long, extending over the end 
of the glass tube. Fill the glass tube with water, then 
invert it, and place the lower end in a dish of mercury, 
having care that the water remains up in the tube far 
enough to cover the end of the branch (Fig. 33). As 
transpiration proceeds, the pressure of the atmosphere 
will force the mercury up the glass tube as rapidly as the 
water passes into the plant. This experiment is some- 
times said to illustrate the " lifting power" of transpira- 
tion, but from the explanation here given, it is seen that 
the mercury is not lifted, but pushed up the tube by the 
pressure of the outside air. This experiment should not 
be regarded as illustrating more that it really does; it 
does not, for example, explain the rise of sap in plants. 
41. Ascent of Sap. It is a well-known fact that, 
although living leaves deprived of water merely become 
wilted, dead leaves eventually dry up; they cannot supply 
themselves with water, although evaporation is taking 
place from their surfaces, and although the stem to which 
they are attached is abundantly supplied. We must 
conclude, therefore, that merely physical forces (imbibi- 
tion and evaporation) are not sufficient to account for 
the rise of liquids in stems. Recent experiments indicate 
that, in this connection, much importance should be 
attached to the secretion of substances by the leaf a 
physiological process. 
We are familiar with such action in the secretion of 
nectar by the nectar-glands of flowers (Fig. 34). Some 
leaves (e.g., Colocasia antiquorum) also secrete water so 
rapidly that it falls in drops from their tips. It is 
probable that, in transpiration, the protoplasm in the 
